The other missing data problem in statistics
Thoughts on improving representation in #StatsTwitter and beyond
Maybe I’m weird, but I really enjoy reading about the history of statistics. There is something about the stories of how statisticians came up with their ideas that makes their work more accessible and memorable for me. In reading this history, I’ve come to appreciate how many of the most important statisticians are largely unknown to scientists, especially those statisticians who made their mark before the latter half of the twentieth century (with a few notable exceptions).
With this in mind, I started the following Sunday #StatsTwitter thread:
#StatsTwitter did not disappoint, and the list of not-as-well-known-as-they-should-be statisticians started to grow. However, and perhaps unsurprisingly, the list was for a few hours composed completely of men. Professor Sherri Rose, a widely respected statistician, was kind enough to point this out.
I am now going to admit to something that some of you might find heretical, but after reading her tweet, I cocked a cynical eyebrow and perhaps even let out a harrumph. After all, the thread was only a few hours old. And it’s not my fault that statistics had such strong British and (later) American influences in the 19th and 20th centuries! And it’s not my fault that the barriers to becoming a statistician in those times and places were much much lower for White men compared to others!! You can’t blame me for the world as it was!!!
And then I took a breath, and remembered that representation matters.
When I see a list of notable statisticians, and they are all White men like me, it doesn’t cause me to wonder about my place in this field (though I have pleny of other insecurites about that I assure you). But if I put myself in someone else’s shoes, it’s not hard to imagine the frustration that might grow from being made to feel like you stick-out or don’t belong in your own profession, especially one you have excelled at, just because you don’t “look the part”. And it’s not hard to imagine the anger that might grow every time someone, maybe even a respected peer, uses flawed, circular reasoning to argue that any lack of representation justifies itself, while ignoring the glaringly obvious structural barriers at the root of the problem. Or to imagine the agitation from hearing, for the upmteenth time, that identity shouldn’t matter so why must you go on about it all the time.
And while things certainly are getting better, this isn’t just a matter of ancient history. Women now make up a large proportion of statistics graduates, but are still disproportionately under-represented in leadership positions; while opportunities in statistics for people who aren’t White are lagging even more, especially for Black people (in the American and UK contexts I am familiar with).
So while I really can’t be blamed me for the world as it was, I can accept some responsibility for the world as it is, and I can take action to encourage a world that I would like to see. One of the easiest ways to do this is to help improve the representation of the diverse community of statisticians that have contributed to our field, even in a casual Sunday #StatsTwitter thread. And while it might be objectively true that the earlier years of our field were, due to circumstance, predominately influenced by White men, it helps many, and hurts none, to be purposefully inclusive when looking back on that history today. Next time I’ll be more careful to do exactly that… and to not assume others will do it for me.
And so my cynical eyebrow dropped back into its normal, slightly-crooked position, and I remembered that Sherri Rose wasn’t attacking me, or scolding me (and she definitely wasn’t attacking or scolding you) - she was sharing her very legitimate frustration (something I feel free to do on the daily, on any number of topics, with little remorse), and simply reminding us that representation does indeed matter, and that we should all feel directly responsible for improving it. And if she drops that reminder with a dose of snark, who I am to take issue with that? It was effective regardless, as people immediately started effortlessly adding the names of women statisticians to the thread and I’m sure I’m not the only one who learned a few new ones (like Elizabeth Gardner and Hilda Mary Woods).
.
On a final note, I want to point out that for the trouble of tweeting this reminder, a handful of people replied to Sherri, complaining that it wasn’t constructive, or that she should offer some suggestions herself. Now, while I can understand why someone might at first think it reasonable request for her to add some female examples, I can also see how irritating (to say the least) that would be when it always falls on you to do it. Beyond this, there were the more intentionally inflammatory or dismissive comments, such as the suggestion that she was being a “snowflake”. Unfortunately this kind of “reply-guy” response is immediately recognizable to anyone paying attention. It is impossible for me not to notice after 10 years on the twitter that a lot of our colleagues would simply prefer to lurk, because engagement can so often lead to a flood of needlessly annoying interactions, or even verbal abuse and physical threats. Why would anyone bother? And so #StatsTwitter gets a bit more homegenous, and we are all worse off for it, as experts with important things to share decide it’s simply not worth it to engage.
(Image courtesy of @Ciaraioch)